Abstract
Background/Aim: Adiponectin protects from metabolic disease and cancer. Accordingly, serum adiponectin was reduced in patients with colorectal cancer (CRC). This hepatoprotective factor was definitely increased in hepatocellular carcinoma (HCC). CRC metastases to the liver are common and the aim of the present study was to evaluate whether serum adiponectin discriminates primary from secondary liver cancers. Materials and Methods: Adiponectin was measured by ELISA in the serum of 36 patients with colorectal liver metastases, 32 patients with HCC and 49 patients without cancer. Results: Serum adiponectin levels were higher in cancer than non-tumor patients. Adiponectin was not related to TNM stage in HCC nor to the levels of serum tumor markers. Moreover, hepatic inflammation and liver fibrosis were not correlated with serum adiponectin levels. Metabolic diseases are associated with low adiponectin and a higher risk of cancer. In HCC, but not in CRC serum, adiponectin was increased in patients with hypertension and hyperuricemia. In this cohort, adiponectin positively correlated with chemerin, an adipokine supposed to contribute to metabolic disturbances. Conclusion: Serum adiponectin cannot discriminate primary from secondary liver tumors.
Adiponectin is a well-studied adipokine initially described by Arita et al. to circulate in very high concentrations and to be reduced in obesity (1, 2). Adiponectin improves insulin sensitivity, and has anti-inflammatory and hepatoprotective properties (3, 4). Insulin resistance, hypertension and dyslipidemia are common comorbidities of obese patients, and therefore, adiponectin reduction has been suggested to contribute to the metabolic syndrome (5). Circulating adiponectin levels were decreased in patients with obesity but were also lower in patients suffering from these complications (6). Hyperuricemia is associated with visceral adiposity, and serum adiponectin was also reduced in these patients (7).
Obesity is a risk factor for the development and progression of different cancers, such as hepatocellular carcinoma (HCC) and colorectal carcinoma (CRC) (8). Comorbidities associated with the metabolic syndrome further contribute to CRC and HCC vulnerability (9, 10). Low grade chronic inflammation and altered adipokine levels are supposed to play a causative role herein (8, 11). Risk of several cancers was indeed higher in patients with reduced adiponectin and this association was most relevant in endometrial, prostate, breast, colon, and gastroesophageal cancers (12). Accordingly, most studies have reported reduced serum adiponectin levels in patients with colorectal adenoma and CRC compared with healthy controls (13). In non-metastatic CRC, circulating adiponectin declined with tumor stage and was further associated with CRC recurrence (14). Protective activities of adiponectin in CRC were confirmed in experimental models where tumors were induced by chronic inflammation or implantation of colon cancer cells (15, 16). Current evidence from clinical and experimental studies support the idea that low adiponectin levels contribute to disease initiation, progression and prognosis of CRC (11, 12).
Beneficial effects of adiponectin were further identified in models of alcoholic and non-alcoholic liver diseases (2, 17, 18). Later on, it was found that serum adiponectin was induced in patients with advanced liver injury (19). This was partly attributed to an impaired biliary secretion of adiponectin in patients with chronic liver diseases (11, 20). High adiponectin levels were furthermore associated with an increased risk to develop HCC (11, 21). In patients with HCC, circulating levels of this adipokine were elevated and, in some cohorts, independently predicted overall survival (11, 21-24).
Experimental studies demonstrated a protective role of adiponectin in liver cirrhosis and HCC (25-27). High adiponectin levels in patients with liver cirrhosis and/or HCC were linked to disease progression (11, 19). Low expression of adiponectin receptors in liver tumors points to a state of adiponectin resistance in these patients (28).
CRC metastases are commonly detected in the liver (29). Here, we hypothesized that patients with colorectal liver metastases have lower serum adiponectin levels compared to patients with HCC. Therefore, adiponectin was measured in the serum of both cohorts.
Materials and Methods
Patient cohorts. Patients and non-tumor controls used in the present study have been described recently in detail (30). Anthropometric and laboratory data of the patients are shown in Table I of this freely accessible paper (30).
The liver of the patients was histologically examined and scored as suggested by Kleiner et al. (31). HCC staging was performed as described (32). Prospective collection of CRC patients' serum was done from January 2012 to June 2015, and of HCC patients' serum from May 2012 to May 2015. Inclusion criteria were: 1) histologically confirmed HCC or CRC metastases and 2) age above 18 years. Exclusion criteria was pregnancy.
Serum of controls without tumors was obtained from January 2008 to June 2008. The cohort included hospitalized patients and outdoor patients. These patients were referred to the Interdisciplinary Ultrasound Department of the Hospital and were described recently (33, 34). The study cohort included 56 patients and serum from 49 patients was available for the present study. Patients with malignancies, inflammatory bowel disease, ascites, hepatobiliary diseases, familial hyperlipidemia, drugs that cause hepatic steatosis, infection with HIV, chronic drug or alcohol abuse, and acute medical conditions with confounding effect on laboratory values were excluded from the study. Aliquots of the sera were stored at −80°C to avoid freeze-thaw cycles. Storage time of serum differed for up to 7 years. This may limit the relevance of the findings of our study. There was, however, no difference in serum adiponectin levels of patients with CRC, which were collected in 2012 (6 patients) and 2015 (10 patients; p=0.18). This indicates that a four-year storage period at −80°C did not grossly diminish serum adiponectin levels. A further study showed that frozen storage of plasma at −30°C for 30 months had no effect on adiponectin levels (35). Thus, the changes in adiponectin levels during seven-year storage may be rather small.
Experiments complied with the guidelines of the charitable state-controlled foundation Human Tissue and Cell Research, Germany. Each patient signed a written informed consent. The study was approved by the ethical committee of the Regensburg University Hospital.
Adiponectin ELISA. Adiponectin ELISA was purchased from R&D Systems (Wiesbaden, Germany) and performed as recommended by the manufacturer. Serum was diluted 1:5,000 fold before analysis.
Laboratory value. Metabolites such as bilirubin, aminotransferases and tumor markers were routinely measured at the Institute for Clinical Chemistry and Laboratory Medicine (University Hospital Regensburg).
Correlations of adiponectin with tumor markers and measures of liver disease in the whole cohort of cancer patients and the subgroups.
Statistics. Results are shown as box plots, which indicate the median values, lower and upper quartiles and the range. The points in the graphs indicate outliers, which are values outside of the inner fences. The stars mark extreme outliers where the values were more than 3-times the height of the boxes. Statistical tests used were Mann-Whitney U-Test, Spearman correlation and one-way Anova with post-hoc Bonferroni (SPSS Statistics 25.0 program). A value of p<0.05 was regarded as significant.
Results
Association of serum adiponectin levels with gender, age and BMI. Adiponectin was measured in serum samples of 49 control patients without any tumors, 32 HCC and 36 CRC patients with liver metastases by ELISA. Controls were patients who came to the hospital mainly due to stomach or epigastric pain, but without any cancers.
As expected, HCC patients had increased aminotransferase activities (36) and higher levels of γ-glutamyltransferase (30). The tumor marker alpha-fetoprotein (AFP) tended to be higher in HCC (p=0.065), whereas carcinoembryonic antigen (CEA) and cancer antigen 19-9 (CA 19-9) were significantly increased in CRC patients (30).
Higher adiponectin levels were detected in female non-tumor (p=0.018) and tumor-bearing patients (Figure 1A). In the three cohorts, adiponectin levels did not correlate with age (data not shown). Negative association with BMI was identified in the non-tumor patients only (r=−0.367, p=0.01). Accordingly, adiponectin was similar in normal-weight, overweight and obese tumor patients (Figure 1B). This was also the case when both cohorts were analyzed separately (data not shown).
Adiponectin (APN) in the serum of patients with hepatocellular carcinoma (HCC), colorectal carcinoma (CRC) and non-tumor patients (Control). (A) APN in the serum of 51 male and 17 female tumor patients. (B) Serum APN stratified for body mass index (BMI) in the tumor patients. There were 24, 32, and 11 patients with a BMI <25, ≥25 and >30 kg/m2, respectively. BMI of 1 patient is not known. (C) APN relative to TNM stage in HCC patients (TNM stage: IA/IB/IIA/IIB/III, number of patients 14/8/7/2/1 per stage) (D) APN in 20 HCC patients without and 12 HCC patients with vascular invasion. (E) APN in 13 CRC patients with and 23 CRC patients without chemotherapy. (F) APN in 32 HCC, 36 CRC and 49 non-tumor controls. **p<0.01, ***p<0.001.
Association of serum adiponectin levels with tumor-node-metastasis (TNM) stage, tumor markers and tumor type. In the HCC group, adiponectin was not correlated with tumor size (r=0.169, p=0.355), grade (r=−0.039, p=0.841) or TNM stage (Figure 1C). Patients with and without vascular invasion had similar serum adiponectin levels (Figure 1D). Serum of CRC patients was collected shortly before hepatic resection of the metastases, whereas the primary tumor was discovered up to six years earlier. Therefore, associations of adiponectin with tumor stage/grade were not calculated.
In the cohort of CRC patients, there were 13 patients who received neoadjuvant chemotherapy before liver resection. Adiponectin did not differ in this cohort (Figure 1E).
Adiponectin did not correlate with the tumor markers AFP, CA 19-9 or CEA (Table I).
Relative to the controls, adiponectin levels were higher in HCC and CRC patients with no differences between the two cohorts of cancer patients (Figure 1F).
Association of serum adiponectin levels with comorbidities. Reduced serum adiponectin levels in patients with type 2 diabetes, hypertension, hypercholesterolaemia and hyperuricaemia are well-described in the literature (6, 7). In the non-tumor cohort, adiponectin was lower in the type 2 diabetic patients (p=0.011), but was not changed in patients with hypercholesterinemia (p=0.587) or hypertension (p=0.488) (data not shown). In the tumor-bearing patients, adiponectin levels were, however, not changed in the 21 patients with type 2 diabetes. Adiponectin was not significantly diminished in the 34 patients with arterial hypertension (p=0.079). Moreover, hypercholesterolaemia (11 patients) and hyperuricaemia (7 patients) were not associated with reduced circulating adiponectin levels (data not shown).
Separate analysis of the cohort of CRC patients revealed similar results; serum adiponectin levels were not changed with comorbidities (6 type 2 diabetic patients, 8 patients with hypercholesterolemia, 16 with hypertension and 3 with hyperuricemia; Figure 2A-C and data not shown). In the cohort of HCC patients, the 15 type 2 diabetes patients and the 3 hypercholesterolaemic patients had adiponectin levels comparable to those without this comorbidities (Figure 2D and data not shown). Serum adiponectin was unexpectedly increased in the 18 patients with arterial hypertension and the 4 hyperuricemic patients (Figure 2E, F).
Serum adiponectin (APN) and comorbidities. (A) APN in 6 colorectal cancer (CRC) patients with and 30 patients without type 2 diabetes. (B) APN in 16 CRC patients with and 20 patients without hypertension. (C) APN in 3 CRC patients with and 33 patients without hyperuricemia. (D) APN in 15 hepatocellular carcinoma (HCC) patients with and 17 patients without type 2 diabetes. (E) APN in 18 HCC patients with and 14 patients without hypertension. (F) APN in 4 HCC patients with and 28 patients without hyperuricemia. *p<0.05.
Association of adiponectin with liver dysfunction. Serum adiponectin did not correlate with levels of alanine aminotransferase, aspartate aminotransferase, γ-glutamyltransferase or prothrombin time in the whole cohort, and when CRC patients and HCC patients were analyzed separately (Table I). A modest negative association with bilirubin was identified in the whole study group and in CRC, but not in HCC patients (Table I).
We furthermore analyzed potential associations of serum adiponectin with histologic liver abnormalities. The histological results in the cohort of HCC patients showed that 16 patients had liver steatosis, 18 patients had predominant inflammatory conditions and 23 patients had liver fibrosis. In the CRC cohort, hepatic steatosis was identified in 17 patients, liver inflammation in 13 patients and liver fibrosis in 17 patients. Prevalence of these histologic features was comparable in both cohorts of tumor patients.
Serum adiponectin was negatively related to steatosis grade (Figure 3A), but was not significantly changed in patients without or with fatty liver (Figure 3B). Serum adiponectin did not correlate with inflammation or fibrosis scores (data not shown). Accordingly, serum adiponectin did not change with the extent of inflammation or fibrosis in the whole study group (Figure 3C, D), and when both cohorts were analyzed seperately (data not shown).
Correlation of adiponectin with serum chemerin. Chemerin is an adipokine, which is induced in obesity and is positively associated with dyslipidemia and hypertension (37, 38). A negative correlation of chemerin and adiponectin was detected in non-diabetic obese patients (39). However, adiponectin did not correlate with chemerin in the non-tumor patients analyzed in the present study (r=0.238, p=0.099). Of note, a positive correlation of serum adiponectin and chemerin was identified in the whole study group of cancer patients. Significant association of these adipokines was seen in HCC but not in CRC patients (Figure 4A-C).
Serum adiponectin (APN) and liver injury. (A) Correlation of APN with steatosis grade. (B) APN in patients stratified for hepatic steatosis (25 patients without steatosis (0), 27 patients with grade 1, 5 patients with grade 2 and 1 patient with grade 3; Steatosis grade was not documented in 10 patients). (C) APN in patients stratified for hepatic inflammation (28 patients had no inflammation (0), 29 patients had grade 1 and 1 patient had grade 2; Data of 9 patients were not known). (D) APN in patients stratified for hepatic fibrosis (28 patients had no fibrosis, 25 patients had grade 1, 1 patient had grade 2 and 14 patients grade 4). Statistical test is not valid for groups with 1 patient.
Discussion
The main findings of our study were the comparable serum adiponectin levels in patients with HCC and colorectal liver metastases. Though serum adiponectin was induced in tumor-bearing patients when compared to non-tumor controls, its levels were not related to systemic tumor markers or HCC staging. The described negative correlations of adiponectin with hypertension and hyperuricemia in non-tumor patients (6, 7) were abolished in CRC patients. In the HCC cohort, patients with these comorbidities indeed had increased adiponectin levels. Furthermore, serum adiponectin levels were not related to histologically evaluated liver injury.
HCC mostly develops in the cirrhotic liver (40), and these patients had high serum adiponectin levels at an advanced disease stage (19). Such an induced adiponectin level was not identified in the tumor patients. Taking into account that the few of the enrolled cirrhotic patients had compensated liver disease this is in accordance with previous studies (19).
HCC did not change serum adiponectin levels in hepatitis C virus infected patients (41). Moreover, cirrhotic and non-cirrhotic HCC patients had serum adiponectin levels similar to those of the correponding controls (23). However, high serum adiponectin predicted the risk of HCC development in patients with chronic hepatitis C (21), and has been associated with worse overall survival in HCC patients with different disease etiologies (42). Moreover, increased adiponectin has been identified in HCC patients in comparison to a cohort of matched non-tumor patients and controls (43). Similarly, adiponectin was induced in the cohort of HCC patients when compared to non-tumor patients in the present study. However, unexpectedly, CRC patients with metastases had adiponectin levels comparable to HCC patients. Though separate studies have described a decline of serum adiponectin in CRC patients (14, 44), a further analysis has shown normalization of serum adiponectin in patients with advanced disease (45). Thus, adiponectin is actually induced in CRC patients with liver metastases. This suggestion has to be confirmed in CRC patients with and without metastatic disease.
Correlation of serum adiponectin (APN) and chemerin. Correlation of APN and chemerin in (A) all tumor patients (B) in colorectal cancer (CRC) and (C) hepatocellular carcinoma (HCC) patients. ***p<0.001.
To our knowledge, serum adiponectin levels have not been compared in patients with primary and secondary liver cancers. Here, similar serum levels were found in CRC patients with liver metastases and HCC.
A major advantage of our study is the analysis of associations of serum adiponectin levels with marked changes in liver histology and clinical data of liver function. Serum adiponectin showed a very modest and negative association with liver steatosis. Such correlations have been described before in different patient cohorts (2, 46). Adiponectin did not, however, correlate with hepatic inflammation or fibrosis. Actually, adiponectin levels were not higher in cirrhotics with compensated disease than patients without or with mild liver fibrosis.
Positive associations of adiponectin with the extent of liver fibrosis and negative correlations with hepatic steatosis have been nevertheless described in hepatitis C infected patients (47, 48). In NAFLD, serum adiponectin has been inversely associated with the severity of liver steatosis and fibrosis (49). Another study has confirmed our results that adiponectin did not correlate with fibrosis stage (50). Thus, associations between serum adiponectin and liver fibrosis stages need possibly also to be discussed in relation to the underlying disease. Modest positive correlations of adiponectin with bilirubin were only identified in CRC patients. This indicates that association of serum adiponectin with this cholestasis marker is modified by disease etiology.
Characteristics of the metabolic status of a patient such as BMI, insulin resistance and comorbidities may further affect circulating adiponectin levels (2, 4). Adiponectin has been found to be reduced in insulin resistance (2). Negative correlations with homeostasis model assessment (HOMA) index have also been identified in patients with chronic liver injury, whereas further studies could not detect any associations with this measure (22, 24, 51). Of note, adiponectin has not been associated with BMI and HOMA index in patients with progressed liver cirrhosis (50, 52). In CRC patients low adiponectin has been linked to insulin resistance (53). In our cohort analyzing patients with HCC and CRC, serum adiponectin levels were not related to BMI and were not altered in type 2 diabetes. Similar studies in cohorts with cancer patients have demonstrated that serum adiponectin did not negatively correlate with BMI, dyslipidemia or insulin resistance (22, 54, 55). Well known negative correlations of adiponectin with BMI and insulin resistance (2, 4) were thus abolished in HCC and CRC patients (22, 54).
Of note, the present analysis showed that adiponectin was actually induced in HCC patients with hypertension and hyperuricemia. Comorbidities are in general and possibly by definition more prevalent in cohorts of tumor patients than healthy controls (9, 10). Elevated adiponectin described in HCC cohorts may thus be related to these complications (22). Unchanged and even increased serum adiponectin levels in cancer patients suffering from these comorbities further showed that associations described in the general population were either lost or even reversed in tumor patients (6, 7).
Chemerin is abundant in adipocytes and hepatocytes and circulating chemerin levels were increased in obese patients (37). Previous studies have reported that circulating chemerin was not changed in HCC serum whereas its levels were induced in CRC patients (19, 56, 57). Our recent study showed comparable chemerin levels in the HCC and CRC patients with liver metastases (30). Here, we demonstrated a positive correlation of serum chemerin with adiponectin in HCC but not in patients with CRC. In non-cancer patients, negative correlations of these two adipokines has been described (58), illustrating the dysregulated balance of these two metabolic active proteins, especially in patients with HCC. In summary, the present study demonstrated comparable serum adiponectin levels in patients with HCC and colorectal liver metastases. Positive associations of serum adiponectin with comorbidities and serum chemerin were identified in HCC but not in CRC patients, suggesting a specific function of this adipokine in the pathogenetic sequence of liver disease.
Acknowledgements
The excellent technical assistance of Elena Underberg is greatly acknowledged.
Footnotes
Authors' Contributions
Conceptualization: C.B., T.S.W., A.K.; methodology: S.F.; software: formal analysis, S.F., C.B.; resources: T.S.W., D.S.; writing - original draft preparation: C.B.; writing - review and editing: C.B., S.F., D.S., A.K., T.S.W.
Funding
This research was partly funded by the German Research Foundation, grant number BU 1141/13-1.
Conflicts of Interest
The Authors report no conflicts of interest regarding this work.
- Received October 15, 2019.
- Revision received November 22, 2019.
- Accepted November 28, 2019.
- Copyright© 2020, International Institute of Anticancer Research (Dr. George J. Delinasios), All rights reserved
References
In this issue
Jump to section
Related Articles
Cited By...
- No citing articles found.